coreutils-9.8

[Douglas R. Reno]

New minor version

[Xi Ruoyao]

i18n patch updated: ​https://www.linuxfromscratch.org/patches/downloads/coreutils/coreutils-9.8-i18n-1.patch

[Xi Ruoyao]

Updated the patch again with the latest Fedora revision. This time they are much faster...

[Bruce Dubbs]

GNU coreutils NEWS

Noteworthy changes in release 9.8 (2025-09-22) [stable]

Bug fixes

  'b2sum' will diagnose --length values that are too big.
  Previously it would have silently assumed 512 for any larger values.
  [bug introduced in coreutils-9.6]
  
  'base32' and 'base64' when decoding will again diagnose partially
  padded data that ends with a newline.
  [bug introduced in coreutils-9.5]
  
  'basenc -d -i' will now strip '=' characters from the input
  in encodings where padding characters are not valid.
  [bug introduced with the basenc program in coreutils-8.31]
  
  'cp -p' had spurious "Operation not supported" failures when
  copying to non-NFS files from NFSv4 files with trivial ACLs. 
  [bug introduced in coreutils-9.6]
  
  'cp --sparse=always' missed some opportunities to create holes.
  That is, although the copies had the correct data, sometimes
  data zeros used extents rather than holes.
  [This bug was present in "the beginning".]
  
  cp missed opportunities to create holes when copying from file
  systems like squashfs that support SEEK_HOLE only trivially.
  [bug introduced in coreutils-9.0]
  
  cp, install, and mv now avoid possible data corruption on
  glibc 2.41 and 2.42 systems when copy_file_range is used with ranges > 2GiB,
  avoiding https://sourceware.org/PR33245
  [bug triggered since coreutils-9.0]
  
  'date' supports specifying multiple named formats with the last taking
  precedence.  Previously multiple specifications would induce an error.
  [bug introduced in coreutils-5.90]
  
  'dd oflag=seek_bytes' no longer mistakenly reports errors when the
  output file exists on GNU/Hurd.
  [bug introduced in coreutils-8.16] 
  
  'fold' no longer exhausts memory when processing large inputs
  with a very large --width argument.
  [This bug was present in "the beginning".]

  "ls --size --block-size=\'k" could misalign output in locales
  with multi-byte thousands grouping characters.
  [This bug was present in "the beginning".]

  'nohup' avoids implementation defined behavior setting umask,
  avoiding a FORTIFY runtime failure on Bionic libc.
  [This bug was present in "the beginning".]

  'od --strings' with '-N' now works correctly.  Previously od might
  write a NUL byte after a heap buffer, or output invalid addresses.
  [These bugs were present in "the beginning".]

  'od -w0' will now issue a diagnostic and exit gracefully.
  Previously it would have aborted.
  [bug introduced in coreutils-9.3]

  'od -w' no longer silently mishandles enormous widths like 3037000500.
  Instead, it either outputs correctly or diagnoses a too-large width.
  [This bug was present in "the beginning".]

  'od +N.' (where N is a decimal number) works again as per POSIX.
  [bug introduced in textutils-2.0]

  'od /dev/null ++0' no longer mistakenly treats the ++0 as an offset.
  [This bug was present in "the beginning".]

  'sort' with key character offsets of SIZE_MAX, could induce
  a read of 1 byte before an allocated heap buffer. For example:
  'sort +0.18446744073709551615R input' on 64 bit systems.
  [bug introduced in coreutils-7.2]

  stdbuf now works on AIX.  Previously it would have been ineffective.
  [bug introduced with the stdbuf program in coreutils-7.5]

  'tail -n NUM' no longer can output more than NUM lines if stdin
  is a largish regular file with a nonzero initial offset, and grows
  while 'tail' is reading it.
  [This bug was present in "the beginning".]

  'tail -f -n +NUM' no longer mishandles NUM values >= UINTMAX_MAX
  when the input is seekable.
  [bug introduced in coreutils-9.6]

  'tail --pid' avoids some unlikely races if the kernel reuses PIDs.
  [bug introduced in coreutils-9.5]

  'tty' now exits with status 4 with a special diagnostic if ttyname
  fails even though standard input is a tty.  Formerly it quietly
  pretended that standard input was not a tty.
  [This bug was present in "the beginning".]

New Features

  basenc supports the --base58 option to encode and decode
  the visually unambiguous Base58 encoding.

  'cksum -a' now supports the 'sha3' argument, to use the SHA3-224,
  SHA3-256, SHA3-384, SHA3-512 message digest algorithms depending on
  the argument passed to the required --length (-l) option.

  'cksum -a' now supports the 'sha2' argument, as a more consistent
  interface than the existing 'sha224', 'sha256', 'sha384', 'sha512'
  arguments, which are now selected with the --length (-l) option.

  'date' now outputs dates in the country's native calendar for the
  Iranian locale (fa_IR) and for the Ethiopian locale (am_ET), and also
  does so more consistently for the Thailand locale (th_TH.UTF-8).

  fold now supports multi-byte characters, honoring their column width.
  Also the --characters (-c) option was added to wrap at a certain
  number of characters, similarly to --bytes in uni-byte locales.

  nproc now honors any cgroup v2 configured CPU quotas,
  which may reduce the effective number of processors available.

  stty supports setting arbitrary baud rates on supported systems,
  like Hurd, Linux with glibc >= 2.42, and some BSDs.
  Also on other systems the full set of supported baud rates
  is determined at build time if possible.

  Commands that support hardware acceleration like cksum and wc
  can now disable this acceleration at runtime through the
  commonly used GLIBC_TUNABLES environment variable.  For example
  to disable the use of AVX512 instructions in cksum, you can:
    export GLIBC_TUNABLES='glibc.cpu.hwcaps=-AVX512F'

Changes to conform better to POSIX.1-2024

  readlink now defaults to being verbose if the POSIXLY_CORRECT
  environment variable is set.

  realpath now supports -E, which specifies the default behavior.
  The corresponding long option is --canonicalize.

  tsort now accepts and ignores -w.

Improvements

  'factor' is now much faster at identifying large prime numbers,
  and significantly faster on composite numbers greater than 2^128.

  fold now exits immediately upon receiving a write error,
  which is significant when reading large / unbounded inputs.

  'seq' is more accurate with large integer start values.
  Previously 'seq 18446744073709551617 inf | head -n1' would
  output the number before the user specified start value.

Build-related

  cksum was not compilable by Apple LLVM 10.0.0 x86-64, which
  lacks support for checking for the VPCLMULQDQ instruction.
  [bug introduced in coreutils-9.6]

[Bruce Dubbs]

Fixed at commit 9e2fa9a05d3

Update to vim-9.1.1806.
Update to iana-etc-20250926.
Update to coreutils-9.8.
Update to expat-2.7.3 (Security release).
Update to linux-6.16.9.
Update to markupsafe-3.0.3.
Update to meson-1.9.1.
Update to openssl-3.5.3.
Update to util-linux-2.41.2.